Performance characteristics of theAgilent 1100 Series capillary LC system using diode-array UV and MSfor detection

Technical Note

Abstract

Capillary LC is the method of choice, when sample volumes are limited

or/and higher sensitivity is essential. A LC capillary/MS combination

fulfills these requirements and is a typical solution in the biopharma-

ceutical research environment. This Technical Note describes the

excellent performance of the Agilent 1100 Series capillary LC system

using a diode-array detection system and an Agilent 1100 Series MSD

Ion Trap as detectors. The system offers robustness, easy set-up and

optimized control and data handling software.

This application hasbeen verified using anAgilent 1200 SeriesLC system, and

showed comparable oreven better performance.

Introduction

The combination of capillary LCusing mass spectrometers and UV diode-array detectors is a technique, which is being more and more applied in bio-analyticalwork to analyze peptides and pro-teins. Furthermore, capillary LC isalso frequently used for the analy-sis of drugs and its metabolites.The reason is that in many cases itis necessary to gain sensitivitywhen sample volume is limited or ifthe sample concentration is verylow. Columns with smaller inter-nal diameter offer the benefit ofincreased sensitivity, since sensi-tivity or limit of detection variesinversely with column diameter.The effect can be understood byscaling down of volumes, in thiscase peak volumes, while keepingthe mass constant. Thus, com-pounds are eluted in smaller vol-ume, resulting in an increase ofpeak concentration, an increase inpeaks height and consequently better signal-to-noise ratio. Table 1indicates the theoretical gain inlimit of detection of lower internaldiameter columns compared to 4.6-mm id columns.

Figure 1 is an example, showing theinfluence of decreased columndiameter on the limit of detection.The same amount of sample wasinjected on two different columns.A comparison was made between amicrobore column with an internaldiameter of 1.0 mm and a capillarycolumn with an id of 0.3 mm. Thelimit of detection is better by afactor 10 to 28 on the 0.3-mm idcapillary for the three peptidesanalyzed.

2

20 40 60

0

20

40

60

80

100

120

140

160

LOD = 47 fmol

LOD = 510 fmol

Time [min]

0 20 40 60

Absorbance[mAU]

0

25

50

75

100

125

150

175

200

0 20 40 60

0

50

100

150

200

250

LOD = 48 fmol

LOD = 492 fmol

28x 10x 10x

0

5 pmol Angiotensin

LOD = 17 fmol 0.3mm id capillary

LOD = 478 fmol 1mm id column

5 pmol Met-Enkephalin10 pmol Met-Lys-Bradykinin

Figure 1 Analysis of three peptides on a 1-mm id column and a capillary column of 0.3-mm id. The limit ofdetection is better by a factor 10 to 28 on the capillary compared to the 1-mm id column.

Table 1 Internal diameter vs. gain in limit of detection

Column id (mm) Factor for increase in sensitivity

4.6 to 1.0 21.24.6 to 0.3 235.04.6 to 0.05 8816.0

Chromatographic conditions

Solvent: A = H2O + 0.01 % formic acidB = ACN + 0.009 % formic acid

Capillary Pump gradient: 5 min 5 % B, 60 min 60 % B, 75 min 60 % BStop time: 75 min; Post time: 15 minFlow: Capillary column 4 µl/min, 1 mm column 30 µl/minColumn: ZORBAX 300SB-C18, 0.3 x 150 mm, 3.5 µm

ZORBAX 300SB-C18, 1.0 x 150 mm, 3.5 µmPre-Column: ZORBAX 300SB-C18, 0.5 x 3.5 mm, 3.5 µmColumn temp.: 25 °CPre-Column switch: 5 minInjection volume: 1 µlDAD: 210/8 nm (ref.: 360/100 nm)

Conditions for the 1-mm id columnFlow: 30 µl/min, 100 µl sensor for the capillary system

Although this improvement inlimit of detection has been knownfor years, capillary LC was not frequently used due to problemsregarding robustness and handlingof the instrumentation. Recentlycapillary instruments, that haveovercome these problems havebecome available commerically.One of the most robust and mosteasy to handle instruments is the Agilent 1100 Series capillaryLC system. Since its introductiontwo years ago, more than 300 systems have been sold worldwideand feedback regarding robust-ness, performance and handling isvery positive. Besides capillary LC for flow rates from 1 to 100 µl/min, nano LC for flow rates<1 µl/min is also gaining more andmore interest for analysis of peptides and proteins using MSand MSn as detection system. TheAgilent Nanoflow ProteomicsSolution covers a flow rate rangefrom 0.01 – 1 µl/min (Agilent publi-cation number 5988-3621EN). Thenanoflow solvent delivery systemcombined with the nanospray nebulizer offers additional sensi-tivity advantages for Agilent massspectrometers.

This Technical Note describes the excellent performance of theAgilent 1100 Series capillary LCsystem using 300- and 500-µm idcolumns. This includes:• precision of retention times, • gradient mixing performance,• system delay volume,• precision of areas with injection

volumes in the nl and µl range, • injection volume linearity,

3

• minimum accessible sample volume,

• carry-over with MS Ion Trap, and• limit of detection using the diode-

array detector, MS Ion Trap.

We also describe the expected per-formance, if the Agilent 1100 Seriescapillary LC system is modified totolerate flow rates above 0.1 ml/minand up to a maximum of 2.5 ml/min.

Instrumentation

The Agilent 1100 Series capillaryLC system can be used in normalmode and capillary mode. In normalmode flow rates above 100 µl/minup to 200 µl/min can be used. In thismode the electronic flow controlis bypassed with a capillary. Inaddition, the 1100 Series capillaryLC system can be modified toeven tolerate flow rates up to 2.5 ml/min. This upgrade involveschange of flow capillaries, changeof UV detector cell and installa-tion of the standard purge valve.In the capillary mode column flowrates from 1 up to 100 µl can beset using two different flow sensors:

• The 20-µl sensor is used for a flow rate range between 1 – 20 µl/min, for columns withinternal diameters < 800 µm

• The 100-µl sensor is used for aflow rate range between 20 up to 100 µl/min, for columns withinternal diameter ≤ 1 mm - 2 mmcolumns can be also used, ifflow rates ≤ 100 µl/min areapplicable.

• The primary flow rate, whichdetermines the flow rate of thepump before the flow splittingpoint can be varied betweenthree flow rate ranges:– low solvent consumption

— 200 to 500 µl/min– medium solvent consumption

— 500 to 800 µl/min– high solvent consumption

— 800 to 1300 µl/min for rapidgradient changes

This flow measurement principleensures highly precise retentiontimes and robustness.

The Agilent 1100 Series capillaryLC system used included:

• Micro vacuum degasser with aninternal volume of 1 ml for fastsolvent changeover.

• High-pressure capillary pumpoffering stable flow rates in thelow µl range.

• Micro well-plate sampler for precise injections in the nl and µl range. A sample coolingdevice was used to avoid sampledecomposition.

• Thermostatted column compartment for stable retentiontimes at any temperaturebetween 10 degrees below ambient and 80 °C.

• Diode-array detector with a 500 nl cell for sensitive analysisof small peak elution volumes.

• MSD/SL Ion Trap with optimizedcapillary sprayer for enhancedsensitivity and MSn capabilities.

• Modification kit to upgrade for2.5 ml/min flow rate.

Results and Discussion

Precision of retention timesFigure 2 shows the analysis ofpeptides. 1.4 µl of a HPLC peptidestandard containing 5 ng/µl eachwere injected. The standardincludes small molecules and molecules above MW= 500. Theprecision of retention times is typically < 0.5 % RSD over the complete range. The delay volumeof the autosampler was bypassedafter the sample had reached thetop of the column. This saves runand cycle time.

Impact of backpressure fluctuations on retention timesThe Agilent 1100 Series capillarypump uses proven microbore highpressure mixing technology – thebest way to generate both preciseshallow as well as rapid gradientsfor screening applications.Flow rates down to 1 µl/min aredelivered using an active splittingmode, the so-called electronicflow control (EFC), which involvesa flow sensor together with anelectromagnetic proportioningvalve (EMPV). The column flow isactively measured and kept stableeven at changing pressures. Thisbecomes important when the column back-pressure increases,due to sample particles which maypartly block column frits (figure 3).

In the flow sensor the columnflow is measured precisely and theflow sensor signal triggers theelectromagnetic proportioningvalve (EMPV), which acts as anactive splitter system. Because the regulation process is based onheat capacities each solvent mixneeds its individual calibrationcurve, which is available from

4

Time [min]20 30 40 50 60 70 80

Absorbance[mAU]

0

20

40

60

80

100

120

140

1 Gly-Tyr2 Val-Tyr-Val3 Met-Enkephalin4 Leu-Enkephalin5 Angiotensin II

1

2

3 4

5

Figure 2 Precision of retention times

Chromatographic conditions

Sample: Sigma HPLC Peptide Standard H-2016Column: 0.3 x 250 mm, ZORBAX 300SB-C18, 5 µm, 300 ÅSolvent A: 0.05 % TFA in waterSolvent B: 0.045 % TFA in acetonitrileMixer: 420 ml standard mixerFlow rate: 200 µL/min main, 5.5 µL/min column flowTemp.: 30°CDetection: 206/10 nm, 450/80 nm,

500 µL / 10 mm DAD flow cellPump piston stroke: 20 µL

Pump gradient0 min, 1 %B

120 min, 31 %B130 min, 85 %B132 min, 85 %B140 min, 1 %B175 min, 1 %B

Injector program for bypassing the autosampler delay volume after sample load onto the columnDrawInjectWait 2 minValve bypassWait 168 minValve mainpass

the ChemStation or the controlmodule for most common solventsand solvents mixtures.

Gradient performance – composition accuracy, precision and mixing noiseTo perform highly reproducibleseparations based on gradientruns it is important that the pumpmixes the solvents accurately andprecisely. To determine the com-position accuracy, precision andmixing noise of the capillarypump, step gradients with waterand a water/acetone tracer wereused. As detector, a diode-array detector was used and samplewavelengths at 265 nm wereselected. A restriction capillaryreplaced the column. The experiments were performed for a flow rate of 4 µl/min.

Different instrument configurationswere tested changing mixer, injectorparameters and the primary flowrate range. For the experiments allinstrument configurations weretested for step height accuracy,step height precision, step mixingnoise and delay volume.

Step height accuracy

The way step height accuracy iscalculated here indicates whetherthe gradient provides the correct% of the binary gradient.Step height accuracy calculation

The height of the 100 % B (tracer)is measured and the chromato-gram is rescaled from 0 % to 100 % B. For each step, the differ-ences between the theoretical and the measured step height is deter-mined as percentage

5

Time [min]0 2 4 6 8 10

Absorbance[mAU]

0

20

40

60

80

100

Pressure 110 bar

Pressure 66 bar

Analysis at 110 bar

Analysis at 66 bar

1

2

3

4

Figure 3 Analysis of phthalates, biphenyl and o-terphenyl at different pressure values: 66 respectively 110 bar.An important criterion is the flow stability for different pressure levels indicated by stable retentiontime data.

Chromatographic conditions

Sample: (1) methylphthalate-, (2) ethylphthalate and (3) biphenyl o-terphenyl (4)

Column: ZORBAX Eclipse SB-C18, 150 x 0.5 mm, 5 µmFlow rate: 0.01 ml/minMobile phases: Water/ Acetonitrile = 30/70Injection volume: 200 nlColumn temperature: 30 °CDiode array detector wavelength: 225/20 nm, ref 550/100 nm

Step height precision

This is an indication whether thegradient formation is the same forseveral experiments performedunder the same conditions.Step height precision calculation

This calculation is based on sever-al step gradient experiments,using the same chromatographicconditions. For each step the stepheight accuracy is determined forat least three runs. From these values the average value and thestandard deviation for each stepare calculated.

Step mixing noise

An indication on how completemobile phases are mixed for anisocratic step. Theoretically, if solvents are mixed 100 %, nonoise should be observed.Mixing noise calculation

The peak-to-peak noise inabsorbance units for several stepsis measured and percentage relatedto a 100 % absorbance step is calculated for each step.

In figure 4 the overlay of threeconsecutive step gradients areshown for 4-µl/min flow rate. Thestep gradient started at 0 % tracer– in 1 % steps the tracer percent-age was increased up to 10 %. Thisis the most critical performancerange is for every pump. The injection valve was switched intothe bypass mode after 1 min toexclude the well plate samplervolume from the system delay volume. It is good practice toshorten cycle times in capillaryLC. The primary flow rate was setto low solvent consumption andthe column flow rate to 4 µl/min. • Accuracy of step height for steps

from 1 to 10 % in 1 % steps hasan average deviation of 0.036 to 0.066 %, close to theoreticalvalues.

• The step height precision for 1 %steps between 0 to 10 % tracertypically shows a standard deviation of 0.024-0.041 %.

• Mixing noise was typically < 0.027 % with autosampler inmainpass, 420 ml mixer installedand low solvent consumption.

• Mixing noise was typically < 0.064 % with autosampler inbypass, no mixer and high solvent consumption

Table 2 provides a short overviewof the parameters measured , calculation procedures and results.

6

Chromatographic conditions

Column flow: 4 µl/min Mobile phase: A = water, B=water + 0.5% acetoneGradient: from 0 to 10 % tracer in 1% steps – each step was held for

30 minutes, backpressure: approx. 100 bar for 4 µl/min Compressibility: 46 x 106 bar for water as mobile phasePump piston stroke: 20 µlRun time: 340 minDAD signal: 267/10nm, ref. 360/100nmSwitch of injection valve in bypass: after approx. 1 min

Time [min]0 50 100 150 200 250 300

Absorbance[mAU]

0

20

40

60

80

100Flow rate 4 µl/minLow solvent consumption

Figure 4 Composition accuracy and precision, mixing noise, overlay of 3 runs. Gradient from 0 to 10 % insteps of 1 %.

Parameter Principle Calculation Resultto be measured of measurement procedure

Accuracy of binary 1 % step height accuray 1. Calculation of deviation Typically the average gradient from tracer experiment from theoretical value deviation for 1 % steps

for 0 to 10 % in 1 % steps in mAU units is in the 0.036 to 2. Calculation of deviation 0.066 % rangein percentage relative to a 100 % step

Gradient precision Standard deviation 1. Measurement of step The standard deviation of step heights in heights accuracy for at for 1 % steps is absorbance units least 3 tracer experiments typically in the range for at least 3 tracer for steps from 0 to 10 % of 0.024 to 0.041% experiments tracer in 1 % steps

2. Standard deviation of the average deviation for at least 3 experiments

Performance of Mixing noise on different Peak-to-peak noise is Mixing noise with mixer isocratic solvent steps of a tracer measured for several typically < 0.027 %mixing experiment from steps in absorbance Mixing noise without

0 to 10 % tracer is units and percentage mixer typicallymeasured to a 100 % absorbance < 0.064 %

step is calculated

Table 2 Performance of high pressure gradient Agilent 1100 Series capillary pump

System delay volumeThe lower the flow rates the moreimportant the system delay volume.For a capillary system small delayvolumes between point of injectionand column head, column end and detector cell are of utmostimportance for low peak dispersion.Also, the detector cell should havea low volume. The delay volumesfor the 1100 Series capillary LC system are 0.08 µl before and afterthe column. The internal volumeof the detector is 500 nl. The com-plete system delay volume is about600 µl.

The delay volume indicates theinternal volume of an LC system,which has to be swept until a gradient change can reach the topof the column.

In the following delay time is calculated instead of delay volume.This is due to the fact that the primary flow is split and exactdelay volume calculation is notpractical.

Delay time calculationA step gradient is programmedand the time needed to see thischange in the detector is measured.The time when the gradientchange is seen by the detectorwas selected to be approximately10 times the noise. For theseexperiments no column is used.

Using gradient analysis the delayvolume between start of gradientand reaching the top of the columnis the key factor for decreasingcycle times and consequently saving labor time. The systemdelay time for the 1100 Series capillary LC system was testedusing tracer experiments with 4- µl/min flow rate and different

instrument configurations (table3). The last configuration is the standard configuration, which isbest suited for lowest carry over.Here the autosampler delay volumestays in the flow path during thecomplete run time and the delaytime is 16 min. This configurationshould be used when highly concentrated samples (mg/ml) areinjected. For all other applicationsswitching the micro well-platesampler in the bypass mode cansignificantly reduce the delay volume. The delay volume of theautosampler is purged with the setcolumn flow after the flow splittingpoint in the electromagnetic pro-portioning valve. Switching theinjection valve into the bypassmode saves 300 µl delay volume ofthe autosampler and is the mainstep to reduce the system delayvolume. The delay time is reducedby 50 %. To further reduce thedelay time the primary flow ratewas increased to its maximum.This means the delay volume infront of the splitting point isflushed more rapidly and the delaytime is down to 5.7 min. To obtainlowest delay times the standardmixer, which is located before thesplitting point of the primary flowwas removed to save 420 µl delayvolume and to reduce the delay

7

time down to 5.4. As a result it canbe stated that one of the mostimportant parameters to reducedelay time is bypassing theautosampler delay volume, whichis situated after the split point ofthe primary flow.

Precision of areasPrecision of peak areas dependsmainly on the precision of injectionvolume and is an important characteristic that influencesquantification. Precisiondeteriorates at low volumes andshould therefore be measuredover the complete injection volumerange. In our example precisionwas tested for different injectionvolumes using diazepam as sample.Different sample concentrationswere used to avoid precision difference due to concentration differences. The chromatographicconditions involved a gradientanalysis using water and acetonitrile(ACN) as mobile phases. The gradient started at 5 % ACN andwas increased to 80 % ACN in 15 min. The column flow rate was 5.5 µl/min and the detectionwavelength was 254/10 nm. A 0.3 × 75 mm ZORBAX 300SB C-8, 3.5 µm column was used. Formore details refer to chromato-graphic conditions of figure 5b.

Instrument configuration Delay time Recommended usage

Bypass, high solvent consumption, 5.4 min Short cycle time is an issue.no 420 µl standard mixer

Bypass, high solvent consumption 5.7 min Short cycle time is an issue, optimum mixing required.

Autosampler in bypass mode, 7.9 min Protein, peptide, drug applications. low solvent consumption

Autosampler in mainpass, 16.0 min Lowest carry over is needed whenlow solvent consumption highly concentrated samples (mg/ml)

are injected.

Table 3 Delay times at different instrument configuration. The column flow rate is 4 µl/min.

The following injection volumeswere injected 6 times each: 0.03 µl,0.05 µl, 0.1 µl, 0.5 µl, 1 µl, 5 µl, 8 µl.

The 1100 Series well-plate microsampler has two different injectionloop capillaries for two differentinjection volume ranges• up to 8 µl and for a practical

injection volume range from 30 nl to 8 µl

• up to 40 µl for a practical injectionvolume range from 200 nl to 40 µl(for precision data see figure 9)

Table 4 summarizes the results fordifferent injection volumes usingthe 8-µl loop capillary.

It is quite obvious that between 30- and 50 nl injection volume precision of areas can typically beexpected < 10 %. For injection volumes below 500 nl the preci-sion of areas is between 2 and 3 %.From 500 nl upwards the preci-sion of areas is typically ≤ 1 %.

Injection volume linearityInjection volume linearity isanother important parameter inchromatography if the injectionvolume has to be varied to compensate for different sampleconcentrations. The 1100 Seriesmicro well plate sampler, as mentioned before, has two practicalinjection volume ranges, from 30 nl to 8000 nl and from 200 nl to40 µl. Since the main topic of this Technical Note is capillary LC, we have evaluated the injectionvolume linearity from 0.1 µl up to 8 µl.

8

Injection volume RSD retention times [%] RSD areas [%] Amount of injectedof diazepam [µl] (n=6 each) (n=6 each) diazepam sample [ng]

0.03 0.07 8.4 0.3390.05 0.14 2.2 0.5650.1 0.1 1.6 1.130.5 0.07 0.3 5.651 0.34 1 1.135 0.1 0.2 5.658 0.06 0.7 0.904

Table 4Precision of areas at different injection volumes

Amount [ng]0 5

Area

0

500

1000

1500

2000

2500

3000

3500

4000

1 23

4

5

6 Relative resolution %: -3.7923e-1

Antipyrine: correlation= 0.99995Phenacetine: correlation= 0.99992Diazepam: correlation= 0.99997

Correlation = 0.99997

Figure 5a Injection volume linearity from 0.1 µl up to 8 µl – as an example the linearity curve for diazepamis shown.

Linearity can be expressed in different ways. In figure 5a the correlation between “Area”and “Amount” is evaluated. The correlation is > 0.99991 forthe three drugs. The other, morecommon method is to evaluate the response factor against theinjection volume and to determine

the range of deviation of theresulting data points (figure 5b). The relative standard deviation ofthe response factors is in therange of 2 % over the tested injec-tion volume range for diazepam.For the other two compounds therelative standard deviation for theresponse factors is in the 4 % range.

Minimum accessible sample volumeEspecially when analyzing proteins or compounds in biological fluids, sample volumemay be limited. In this case it isadvantageous, if the micro well-plate sampler is able to drawnearly the complete volume ofvery low sample volumes. Onevaluable feature in this case is theability to find the bottom of a vialor a well. Having found the well orvial bottom the draw position ofthe needle should be adjustable to have access even to very smallamounts of sample. For our experiments we chose 2 differentvial types: a 300-µl conical plasticvial and a 200-µl conical plasticinsert with polymer feet for 2-mlglass vials. These polymer feet actas shock absorbers. This allowsnearly all sample volume to bewithdrawn. For these inserts thedefault settings of the 1100 Seriesmicro well-plate sampler mightnot be appropriate. To be sure thatthe draw position of the needle is correct, it is recommended to testit with a standard solution withknown concentration. Otherwise,it may happen that the needle istoo close to the bottom and theselected injection volume is notdrawn completely. For the microplastic vial the “well/vial bottomsensing” function should be selected and the draw position for the needle should be set to -0.8 mm. Using plastic vials is of advantage if peptide and/or proteins are analyzed, becausethese compounds tend to adsorbon glass.

9

2.23E-03 2.30E-03 2.24E-03 2.28E-03

1.00E-03

1.50E-03

2.00E-03

2.50E-03

3.00E-03

3.50E-03

0.1 0.3 0.5 1 5 8Injection Volume [µl]

2.28E-03 2.28E-03

Response Factor[Amt/Area]

Figure 5b Injection volume linearity evaluated based on response factors

Chromatographic conditions

Compounds: antipyrine, phenacetine, diazepamColumn: 75 x 0.3 mm id ZORBAX 300 SB C-8, 3.5 µmColumn flow rate: 5.5 µl/minPrimary flow rate: 200 to 500 µl/min Mobile phase: A = water and B = acetonitrileGradient: at 0 min 5 % B, at 3 min 5 % B, at 6 min 40 % B, at 7 min 50 % B, at 15 min 80 % BRun time: 20 minPost time: 18 minDAD signal: 254/10 nm, ref. 350/80 nm

Figure 6a shows the analysis ofthree drugs. A 200-µl plastic insertfor 2-ml vials was filled with 10-µlof sample and five 2-µl injectionswere performed. The results showthat for the next full 2-µl injection4 ml must be left in the plasticinsert.

Nearly the complete amount ofsample volume is accessible if thecorrect injector parameters areused. The optimum draw positionin this case was 0.9 mm. In thesecond experiment (figure 6 b) the300-µl plastic micro vial was filledwith 50-µl sample and 5 µl wereinjected 10 times.

For the next full 5-µl injection 10 µl must be available. For chro-matographic conditions refer tofigure 6a (except injector parame-ters). The injection volume was 5 µl, the draw position -0.8 mmwith well/vial bottom sensing.

Carry over using MSD Ion TrapCarry over is one of the severeproblems that can falsify quantitative and qualitative data in HPLC. Today detectors arebecoming more and more sensitiveand quantitation in the low pico-gram range is achievable especiallyusing highly sensitive MS detection.When samples of very high (mg/ml)and very low (ng/ml) concentrationare analyzed in a row, carry overbecomes a critical parameter. Werecommend using the 1100 Seriesmicro well-plate sampler forobtaining lowest carry over.Having optimized this sampler,carry over of approximately 0.01 %can be expected even for criticalcompounds (see following applica-tion example). Optimization forlowest carry over involves clean-

10

Time [min]0 20 40 60 80

Absorbance[mAU]

0

200

400

600

800

1000

1200

1400

Air-peak1

2

3

Figure 6a Minimum accessible sample amount from a 200 µl conical plastic inserts. 5 times 2 µl were injectedfrom a 10 µl sample.

Time [min]25 50 75 100 125 150 175

Absorbance[mAU]

0

100

200

300

400

500

Air-peak

12

3

Figure 6b Minimum accessible sample amount from 300 µl conical plastic vials

Chromatographic conditions

Compounds: antipyrine (1), phenacetine (2), diazepam (3)Column: 75 x 0.3 mm id ZORBAX 300 SB C-8, 3.5 µmColumn flow rate: 5.5 µl/minPrimary flow rate: 200 to 500 µl/min Mobile phase: A = water and B = acetonitrileGradient: at 0 min 5 % B, at 3min 5 % B, at 6 min 40 % B, at 7 min 50 % B, at 15 min 80 % BRun time: 20 minPost time: 18 minDAD signal: 254/10 nm, ref. 350/80 nmInjection volume: 2 µl, draw position 0.9 mm, no well/vial bottom sensing, 200 ml polypropylene

inserts with polymer feet for 2-ml glass vials

ing of the needle exterior andcleaning of the injection valvegrooves by switching the valve twoor three times after all peaks areeluted. We used azithromycine inour experiments, which is a verylarge and complex molecule. It isused as wide-band antibiotic. Ithas a large number of functionalgroups, which enable the moleculeto adsorb at different surfaces.The molecule also shows a cagecharacter, which means it cancatch metal ions in its ring follow-ing certain transportation process-es. All these characteristics are ahigh risk for carry over of thismolecule

To be able to analyze azithromycinewith DAD and MS Ion Trap themobile phase had to be basic andvolatile. An isocratic analysis wasapplied using 85 % methanol and15 % of an ammonium acetatebuffer at pH = 9. The capillary column had an internal diameterof 0.5 mm. The concentration ofthe injected azithromycine samplewas as high as 1 µg/µl and 2 micro-liter were injected. The wash solventfor the exterior of the injectionneedle was methanol/water = 1/1,because the molecule dissolves inthis solvent mixture. To test carryover the highly concentrated sample was injected, followed bythe injection of the same amountof mobile phase solution, and followed by a blank run. Anotherimportant factor for obtaininglowest carry over, is not to useoverlapped injection mode or thebypass mode for the injectionvalve when highly concentratedsamples are injected. Highly

concentrated means samples in the high µg to mg/ml range. For such a demanding sample asazithromycine it is not sufficient toclean the exterior of the needle–the injection valve grooves alsohave to be cleaned. This cleaningis done using an injector program,shown here. The valve is switchedseveral times after the peak(s) ofinterest has passed the detector.In addition to external needlewash, this ensures lowest carryover also for critical compounds.

Figures 7a and 7b show the analysisof azithromycine. Figure 7a showsan overlay of the 2-µl injection of the2 mg/l solution followed by injecting2 µl of mobile phase with optimizedcleaning conditions. In this case thecarry over is about 0.01 %. For theMS analysis of azithromycine a sim-ple MS ion trap method was used.Here it is most important to exactlyspecify the target mass, in order toobtain maximum sensitivity.Azithromycine is a doubly positivelycharged molecule with a mass of375.

11

Intensity

2

4

6

8

x108

Base peak chromatogram

0

2

4

x106

5 10 15 20 25 30 35Time [min]

0

Sample injection 2 µg azithromycine

0.01 % carry-over in blind solvent injection with 10 sec. needle wash and injector program

Figure 7aAnalysis of azithromycine and evaluation of carry over without any wash cycles

O

OCH3

CH3

H3C

HO

HO

H3C

H3C

H3CN

CH3CH3

CH3

CH3

CH3 CH3

CH3

CH3

O

O O

HOOH

HOO

O

NN

Azithromycine

Figure 7b shows how the differentinstrument setting influence thepercentage of carry over.

Intensive examination showedthat using overlapped injectionmode or bypass mode, carry overwas extremely high with 0.95 %(figure7b). This is due to the factthat in this mode the completeinjection system is not flushedduring the run time. Withoutbypassing the autosampler volumecarry over is already down to 0.16 %. Washing the exterior of the injection needle reduces the carry over down to 0.04 %. Additional injection valve switchesafter elution of the peak of interestreduces the carry over further,down to 0.01 %. Even thoughazithromycine is a very difficultsample regarding carry over, the1100 Series capillary LC systemwith the micro well-pate sampleris able to provide a carry over aslow as 0.01 %, if some precautionsare taken. These include:• Washing the needle exterior

with a solvent that is able to dissolve the sample.

• Using an injector program to wash the injection valvegrooves after the peak of interest has passed the detector

• Avoid using the bypass mode for the injection valve to reducethe delay volume of the autosampler over the complete run time.

12

06

x106

Intensity

04

x106

0

2x107

0

x106

0

4x106

5 10 15 20 25 30 35Time [min]

4

Base peak chromatograms

0.01 % carry-over in blind solvent injection with 10 sec. needle wash and injector

0.04 % carry-over in blind solvent injection with 30 sec. needle wash

0.08 % carry-over in blind solvent injection with 10 sec. needle wash

0.16 % carry-over in blind solvent injection without needle wash

0.95 % carry-over in blind solvent injection in bypass mode

Figure 7b MS analysis of azithromycine and evaluation of carry over using different instrument set ups anddifferent cleaning methods

Chromatographic conditions

Solvent: 85 % MeOH, 15 % 5 mM NH4Ac,pH 9 with NH4OH

Stop time: 40 minFlow: 5 µl/minColumn: ZORBAX SB-C18,

0.5 x 150 mm, 5.0 µmColumn temp.: ambientInjection volume: 2 µlConcentration of azithromycin: 1 mg/mlWash solvent for needle exterior: MeOH/water 1/1DAD: 210/8 nm (ref.: 360/100 nm)Injection sequence: 1. sample, 2. solvent,

3. blank run

Injector program for lowest carry over:Draw def. amount from sampleNeedle wash as methodInjectWait 35 min.Valve bypassWait 0.2 min.Valve main pathWait 0.2 min.Valve bypassWait 0.2 min.Valve main path

MS conditions

Source: ESIDrying gas flow: 10 l/minNebulizer pressure: 20 psiDrying gas temperature: 325 °CICC: OnAverages: 5Max accu time: 300 msTarget: 20,000Ion mode: positiveTarget mass: 375 [M+2H]2+

Limit of detection using DAD, MSDand MSD TrapWhenever the compounds of interest are available in very low concentrations or only lowsample volumes are accessible,the sensitivity of the detector is oneof the most important factors thatinfluences accurate measurement.Nowadays, UV detection, MSquadrupol and MS Ion Trap detec-tors are able to determine com-pounds in the low femto-molerange. In the following the limit ofdetection for three peptides wastested using the Agilent 1100Series MSD Ion Trap and the Agi-lent 1100 Series diode-array detec-tor. For these experiments a capil-lary column with an internal dia-meter of 0.3 mm was used. Themobile phase was water and ace-tonitrile with formic acid as modi-fier instead of TFA (trifluoro-acetic acid). TFA is not ideal whenusing MS detection because it sup-presses the signal by forming ionpairs with peptides and proteins.Therefore, formic acid was usedwhich shows better compatibilitywith mass spectrometers.

The MS conditions were set suchthat the target mass of the expectedpeptide mass was selected. Theion mode was selected to be positive. These parameter settings ensure that all differentpeptide masses were recognized.One analyzed peptide was MetEnkephalin, a relative short peptide with an average mass of573.7. It forms a positive ion in the MS and is analyzed as singlecharged molecule with a mass of574. The next peptide wasAngiotensinII with an average massof 1042.20. It forms a doubly charged

13

Extracted Ion Chromatogram

5 pmol Angiotensin 5 pmol Met-Enkephalin

0 20 40 60Time [min]

x107

0.0

0.5

1.0

1.5

2.0

2.5

Intensity

LOD =17 fmol

0 20 40 600

1

2

3

LOD = 15 fmol

0 20 40 600

2

4

6

LOD = 31 fmol

10 pmolMet-Lys-Bradykinin

Figure 8a LODs MSD Ion trap extracted ion chromatogram

20 40 60

0

20

40

60

80

100

120

140

160LOD = 47 fmol

Time [min]0 20 40 60

Absorbance[mAU]

0

25

50

75

100

125

150

175

200LOD = 17 fmol

0 20 40 60

0

50

100

150

200

250LOD = 48 fmol

5 pmol Angiotensin II 5 pmol Met-Enkephalin

0

10 pmol Met-Lys-Bradykinin

Figure 8b LODs DAD

Chromatographic conditions

Solvent: A = H2O + 0.01% FA; B = ACN + 0.009 % formic acid

Capillary pump gradient: 5 min 5 % B, 60 min 60 % B,

75 min 60 % BStop time: 75 minPost time: 15 minFlow: 4 µl/min Column: ZORBAX 300SB-C18,

0.3 x 150 mm, 3.5 µmPre-column: ZORBAX 300SB-C18,

0.5 x 3.5 mm, 3.5 µmColumn temp.: 25 °CColumn switch: 5 minInjection volume: 1 µlDAD: 210/8 nm (ref.: 360/100 nm)

positive ion with a mass of 523.8.The last peptide that was analyzedwas Met-Lys-Bradykinin with anaverage mass of 1390.60. This molecule forms a triply chargedpositive ion with a mass of 440.8.

This comparison confirms thattypically the MSD Ion Trap canachieve lower limits of detectioncompared to a UV detection system.The difference in sensitivity willdepend on chromophore properties,ionization properties and others.Another significant advantage of theMSD Ion Trap is the identification of peptides by their masses. Also,with the selected ion mode a peptide can be selectively searched and identified among other co-eluting compounds.

Performance of 1100 Series capillaryLC system modified for flow ratesabove 0.1 ml/minThe Agilent 1100 Series capillary LCsystem can be modified to toleratehigher flow rates. This offers theflexibility of both a capillary LC system and an instrument, which canbe used for standard applicationswith 2.1 and 4.6-mm id columns.

For column ids up to 2 mm andflow rates up to 0.2 ml/min themodification involves bypassing of the electronic flow controlusing a capillary. If flow ratesabove 0.2 ml/min are needed themodification involves:• bypass of the electronic flow

control,• installation of the standard

purge valve,• change of UV detector cell, and• change of flow path capillaries

with small internal diameters tocapillaries with bigger internaldiameters.

14

MS conditions

Source: ESIDrying gas flow: 10 l/minNebulizer: 20 psiDrying gas temperature: 325 °CSkim 1: 40 VCap exit offset: 135 VICC: OnAverages: 5Max accu time: 300 msTarget: 20.000Ion mode: positive

Column (id and length) Flow rate (ml/min) Pressure range

100 x 2.1 mm 0.4 38 to 92 bar0.8 68 to 174 bar

125 x 4 mm 1 45 to 131 bar1.5 67 to 190 bar

100 x 4.6 mm 2 86 to 213 bar2.5 112 to 272 bar

6 port micro switching valve with pre-column 1 76 to 192 barand analytical column in series6 port micro switching valve with pre-column 2 347 bar at 50/50and analytical column in series (water/acetonitrile)

Furthermore, in the Agilent Chem-Station software, the setup for theloop size has to be changed andthe draw speed for sampling hasto be increased to avoid long waittimes until a sample is drawn up.After this modification the follow-ing pressures drops for different column ids and flow rates can be expected (table 5).Water andacetonitrile were used as mobilephase. The gradient applied start-ed with 100 % water and endedwith 100 % acetronitrile. The pres-sure drop was tested at each 10 %step. If more viscous mobile phaseare used the flow rate range maybe reduced.

In some cases the 1100 Series capillary LC system may beequipped with a micro-switchingvalve, which is installed in the column compartment. The capillaries connected to this valvehave to be replaced by 100-µmcapillaries. The pressure drop afterthe modification was measured foran analytical and a pre-columnswitched in series is as follows. Ata flow rate of 1 ml/min the pressuredrop for 100 % water was found tobe 192 bar and for 100 % acetonitrile76 bar. At 2 ml/min and water/ acetonitrile of 50/50 the pressuredrop was 347 bar for the previouslydescribed column configuration.

Table 5 Pressure drop at different flow rates and different column ids, using water and acetonitrile from100 % water to 100 % Acetonitrile

The performance of the modifiedsystem was tested for:• precision of retention times and

areas, • delay volume,• mixing performance, and • step height differences for 10 %

steps.

The precision of retention timesand areas was tested using a mixture of 3 compounds andinjection volumes between 1 and40 µl (figure 9). The precision forretention times was found to be < 0.1 % RSD and for areas <0.5 %over the complete injection range.

The delay volume, the mixing performance and the step heightdifferences for 10 % steps wastested using the tracer experimentin figure 10.

The delay volume depends on themixer. Using the 1100 Series standard mixer the delay volumewas found to be 620 µl. Using thelow volume Upchurch mixer thedelay volume could be reduced to420 µl. The mixing performance isexcellent for all evaluated 10 %steps. The step height varies overten 10 % steps by 2.14 % RSD.

15

Time [min]0 1 2 3 4 5 6

Absorbance[mAU]

0

20

40

60

80

100

120

140

Phenacetine

Diazepam

Antipyrine

Compound RSD RT (%) RSD Area (%)Antipyrine <0.1 <0.5Phenacetine <0.1 <0.5Diazepam <0.1 <0.5

Figure 9 Precision example of retention times and areas for an injection volume of 5 µl

Time [min]0 10 20 30 40 50

Absorbance[mAU]

0

100

200

300

400

500

600

Figure 10 Tracer experiment for testing delay volume, mixing performance and step height precision

Chromatographic conditions

Sample: antipyrine, phenacetine and diazepamColumn: ZORBAX Eclipse XDB-C8, 4.6 x 150 mm, 5 µmFlow rate: 1.2 ml/minMobile phases: water/ acetonitrile = 70/30Gradient: at 0 min 30 % ACN, at 5 min 80 % ACN, at 5.5 min 30 % ACN,

at 7 min 30 % ACNInjection volume: 5 µl Column Temperature: 40 °CDiode array detector wavelength: 254/20 nm, ref 350/80 nm

Results and conditions

System delay approx. ALS mainpass 620 µlvolume: with standard mixer with

Upchurch mixer approx. 420 µl at 170 bar

Step height from 0 to 100 % indifferences: steps of 10 % = 2.14 %Ripple 20 % tracer = 0.032 %Ripple 50 % tracer = 0.004 %Ripple 90 % tracer = 0.006 %Mobile phases: water/water + 0.5 % acetoneFlow rate: 1 ml/minStep gradient: from 0 to 100 % tracer in

steps of 10 %, each step was held for 5 min

DAD: 265/10, ref 350/80

© 2002 - 2007 Agilent Technologies

Printed May 1, 2007Publication Number 5988-7511EN

www.agilent.com/chem/1200

Summary

Capillary LC is the method ofchoice, when sample volumes are limited or/and higher sensitivityis essential. A LC capillary/MScombination fulfills these demands

• RSD of retention times for sigma peptide standard: < 0.5 % RSD

• System delay volume: • 5 µl from electronic flow control (EFC) to column head with 20-µl flow sensor, • 14 µl for 100 µl flow sensor,

80 nl from point of injection to column and from column end to detector

• RSD of areas between 30 and 50 nl injections: < 10 %

• RSD areas for <500 nl injections: 2 to 3 %

• RSD of areas for >500 nl injections: ≤ 1 %

• Carry over: • < 0.01 % for caffeine with external needle wash on the micro well plate sampler; • 0.01 % for azithromycine if special injector parameter set up is used

• Minimum sample volume: approx. 2 µl from 4 µl using 100 µl vials and if applicable well/vial bottom sensing and drawposition – 0.8 mm

• Limit of detection for peptide standard using DAD: approx. 50 fmol

• Limit of detection for MS-Ion Trap with capillary sprayer: < 50 fmol

and is a typical solution in the biopharmaceutical research environment. The Agilent 1100Series capillary LC system offersrobustness, easy set-up and optimized control and data handling software. Detectors

used are the diode array detectionsystem and a MSD Ion Trap. Typi-cally, the 1100 Series capillary LCsystem offers the following perfor-mance based on previouslydescribed experiments: